Abstract: In an open real-time system environment, the coexistence of different kinds of real-time and non real-time applications makes the system scheduling mechanism face new requirements and challenges. One two-level scheduling scheme of the open real-time systems is introduced, and points out that hard and soft real-time applications are scheduled non-distinctively as the same type real-time applications, the Quality of Service (QoS) cannot be guaranteed. It has two flaws: The first, it can not differentiate scheduling priorities of hard and soft real-time applications, that is to say, it neglects characteristic differences between hard real-time applications and soft ones, so it does not suit a more complex real-time environment. The second, the worst case execution time of soft real-time applications cannot be predicted exactly, so it is not worth while to cost much spending in order to assure all soft real-time applications not to miss their deadlines, and doing that may cause resource wasting. In order to solve this problem, a novel two-level real-time scheduling mechanism (including scheduling profile and scheduling algorithm) which adds the process of dealing with soft real-time applications is proposed. Finally, we verify real-time scheduling mechanism from two aspects of theory and experiment. The results indicate that our scheduling mechanism can achieve the following objectives. (1) It can reflect the difference of priority when scheduling hard and soft real-time applications. (2) It can ensure schedulability of hard real-time applications, that is, their rate of missing deadline is 0. (3) The overall rate of missing deadline of soft real-time applications can be less than 1. (4) The deadline of a non-real-time application is not set, whereas the scheduling algorithm that server 0 S uses can avoid the “starvation" of jobs and increase QOS. By doing that, our scheduling mechanism is more compatible with different types of applications and it will be applied more widely.
Abstract: Advances in computing applications in recent years
have prompted the demand for more flexible scheduling models for
QoS demand. Moreover, in practical applications, partly violated
temporal constraints can be tolerated if the violation meets certain
distribution. So we need extend the traditional Liu and Lanland model
to adapt to these circumstances. There are two extensions, which are
the (m, k)-firm model and Window-Constrained model. This paper
researches on weakly hard real-time constraints and their combination
to support QoS. The fact that a practical application can tolerate some
violations of temporal constraint under certain distribution is
employed to support adaptive QoS on the open real-time system. The
experiment results show these approaches are effective compared to
traditional scheduling algorithms.